1. Introduction
Economic theory suggests that given small income effects and available substitutes, an
individual’s willingness to pay (WTP) for a good will be roughly equivalent to his or her
willingness to accept compensation (WTA) for the good (Hanemann, 1991; Randall and Stoll,
1980; Willig, 1976). Despite this theoretical expectation, a significant disparity between WTA
and WTP is usually observed in lab experiments (Horowitz and McConnell, 2002).1 A common
explanation for the disparity, the endowment effect, suggests that an asymmetry exists in
individual preferences and that valuations depend, in part, on the perceived ownership of the
good. If the endowment effect were found to be a fundamental aspect of individual preferences,
it would undermine the validity of the commonly used (Hicksian) model of consumer
preferences that underpins the use of benefit cost analysis to inform public policy (Knetsch,
2010).
Not all WTA-WTP lab experiments report a significant disparity. Plott and Zeiler (2005)
show that procedural controls and subject misconceptions concerning those controls can have
substantial influence on subjects’ behavior. Their experimental procedures were designed to
control for subject misconceptions and were, by and large, motivated by concerns raised in the
literature. These include subject anonymity, an incentive compatible elicitation mechanism, and
explicit training and practice. Their research suggests that the value disparity can be turned on
and off in the lab using a set of procedural controls and, thus, may not be a fundamental aspect of
preferences.2
While the Plott and Zeiler (2005) procedures represent a considerable contribution to our
understanding of the value disparity, their procedures include an important step that is not
motivated by the literature. After initial offers were recorded in the auction of interest, subjects
were prompted to consider whether the offer chosen was their actual non-strategic value and
were then allowed to revise their offer before committing to it. Being unaware of any theoretical
or empirical motivation for including such a prompt, and having found no discussion of prompts
in the literature, we sought to measure the impact of the prompt and revision opportunity on
offers elicited in a carefully implemented WTA-WTP disparity experiment. Like Plott and Zeiler
we are unable to reject equality between WTA and WTP based on the post-prompt and revision
offers. However, using the pre-prompt and revision offers the data reveal a significant value
disparity. Results indicate that the prompt and revision opportunity here has a substantial impact,
systematically increasing WTP and decreasing WTA offers.
2. Experimental Design and Results
Plott and Zeiler (2005) hypothesized that the observed WTA-WTP disparity was the
result of subjects’ misunderstanding of the incentives implicit in the bidding mechanism they
were instructed to use, or of unintended cues given to subjects. Using what they term a “revealed
theory methodology” that accounts for all identified sources of subject misconception, they
1
Whether behavior observed in a lab is of interest to economists who are primarily interested in how individuals
behave within functioning markets is a relevant question first posed in Binmore (1999).
2
Using similar procedural controls in an experiment valuing lotteries, Isoni et al. (2011) found that the disparity
persisted, suggesting that the disparity is sensitive to the nature of the good being traded.
designed an experiment to control for these concerns. These procedures included the following:
subject anonymity to assure that the subjects ignore the social implications of their choice, such
as a desire to impress others; use of an incentive-compatible random price auction to elicit actual
values; and explicit training procedures to ensure that subjects understand that stating offers that
represent their actual value is the optimal strategy when participating in a random price auction.3
Importantly, and of particular interest here, during the binding round all subjects recorded an
initial offer, either a WTA or a WTP depending on their assigned role, and were then prompted
to “consider whether the offer chosen was the actual non-strategic value” and were allowed to
change the offer before committing to it.
The experimental design used here incorporates many of the procedures implemented by
Plott and Zeiler in their Experiment 2, although we do not completely replicate the analysis of
Plott and Zeiler (2005).4 Each subject received a $10 show up fee and was given a mug whose
(unannounced) market value was $7.95. Each subject was also handed an instruction packet and
a recording packet to ensure anonymity. The instruction packet explained the procedures and
provided explicit training with the elicitation mechanism (an open-ended Becker-DegrootMarschak random price auction)5. This training included numeric examples and explanations
illustrating that it is in the subjects’ best interest to state their actual valuations during the binding
mug round. Like Plott and Zeiler (2005), our design suggested to the subjects how they should
determine their individual valuations. Subjects were encouraged to use an iterative valuation
approach wherein the subjects begin the determination of their WTP (WTA) with some low
(high) dollar amount and then ask themselves whether they would be willing to buy (sell) the
item at this amount. If the answer is yes they are then to consider increasing (decreasing) that
amount. This process is to continue until subjects come to a dollar amount that makes them
indifferent between buying (selling) the item and the dollar amount. Thus, each subject was
instructed to iteratively increase their WTP or decrease their WTA upon introspection of their
respective values.
To ensure understanding of the procedures and to demonstrate how the auction would
take place, each subject participated in practice buying and selling tasks [Rounds 1 and 2]. The
items in each practice round had determined values such that subjects were effectively buying
and selling dollar amounts. Finally, in the paid mug round [Round 3] each subject acted as either
a buyer or seller depending on the pre-determined assignment indicated in their recording
packets. After offers were collected all subjects were asked “Before you commit to your Round 3
offer and we reveal the fixed offer, consider whether your offer reflects your actual, nonstrategic, value of the mug. Round 3 will be discarded”. All subjects were required to resubmit
3
The set of experiments performed by Plott and Zeiler also included paid lottery rounds to provide additional
practice with the elicitation mechanism. However, it is demonstrated, in their Experiment 2, that this paid practice
is not essential to their results as it took place after the paid mug round of interest.
4
Our experiment differed from Plott and Zeiler’s Experiment 2 in the following ways: (1) some minor differences in
wording and values used in the instructions, (2) we used a larger show up fee ($10 rather than $5), (3) our mug had
a slightly higher (unannounced) market value ($8.50 as opposed to $7.95), (4) our binding mug round was not
followed by paid lottery rounds, and (5) in our design each subject was given a random ID number but was paid,
anonymously, after the experiment by the experimenter, whereas in the Plott and Zeiler design each subject was
given a random ID card and instructed to bring this card to a third party who simply exchanged each ID card for an
envelope containing the subject’s earnings.
5
Exact experimental instructions and materials are available upon request from the authors.
an offer on Round 4, although they were informed that they were not required to alter their
offers.
Table 1. Individual Offers and Descriptive Statistics by Conditiona
Mean
Median
Std. Dev.
2.59
2.00
2.06
4.87
5.00
3.02
Pre-Prompt
WTP (n=26)
WTA (n=27)
0, 0, 0, 0, .01, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2,
3, 3.5, 3.5, 3.5, 4, 4, 4.25, 5, 6, 6, 7.5
0, 2, 2, 2, 2.4, 2.5, 3, 3, 3, 3, 3.5, 4, 4, 5,
5, 5, 5, 5, 5, 6, 6, 6.99, 7, 8, 8, 10, 15
Post-Prompt
0, 0, 4, 0, 2, 1, 2, 4, 2, 2, 3, 2, 2, 2, 3.25,
3.12
3.00
1.98
2, 4.5, 3, 4, 4.5, 4, 4.3, 5, 7, 6, 7.5
0, 2, 2, 3, 2.4, 2, 2, 2, 3, 3, 3.5, 4, 2, 4,
WTA (n=27)
4.22
4.00
2.25
5, 5, 5, 5, 5, 6, 5, 6, 5, 8, 6, 8, 10
a
Pre-prompt (Round 3) offers are listed in ascending order. This initial order is
maintained for the post-prompt (Round 4) offers.
WTP (n=26)
A total of 53 subjects participated in the experiment, with 26 randomly assigned as a
buyer and 27 as a seller. The individual pre- and post-prompt offers are listed in order of
ascending pre-prompt offers in Table 1. Note that subject order is maintained across pre- and
post-reporting such that the third WTP offer recorded indicates that this subject initially offered
$0 but revised his or her offer to $4 when given the opportunity. The Round 3 (pre-prompt)
offers yield a mean (median) WTP of $2.59 ($2) and WTA of $4.87 ($5). In comparison, the
Round 4 (post-prompt) offers yield a mean (median) WTP of $3.12 ($3) and WTA of $4.22
($4).6 The Wilcoxon Rank-Sum and Median tests are used to test the hypothesis of equality
between WTP and WTA. Like Plott and Zeiler (2005), we are unable to reject equality between
WTP and WTA of the post-prompt offers (Table 2). However, using the pre-prompt offers, the
data indicate a statistically significant value disparity. Further, from Round 3 to Round 4, 11 of
the 26 WTP offers increased while only two decreased and 11 of the 27 WTA offers decreased
while only one increased. Thus, 25 of the 53 subjects (47%) revised their offers, with 88% of
these revisions in the direction implied by the iterative valuation approach.
Table 2. Results
Wilcoxon Rank-Sum Test
(null hypothesis: identical distributions)
z
p-value Conclusion (𝛼 = .05)
Pre-Prompt 3.08 .0021
Reject null
Post-Prompt 1.9
.06
Cannot reject null
a
Pearson 𝜒 � statistics are corrected for continuity.
6
Median Testa
(null hypothesis: populations have identical medians)
p-value
Conclusion (𝛼 = .05)
Pearson 𝜒 �
4.4
.036
Reject null
.48
.49
Cannot reject null
For reference Plott and Zeiler, Experiment 2, post-prompt results were reported as follows: among the 14
subjects assigned to the WTP condition, mean (median) offers were $7.88 ($6.50); and among the 12 subjects in
the WTA condition, mean (median) offers were $5.71 ($5.10). Using both a Wilcoxon Rank-Sum test and a Median
test Plott and Zeiler are unable to reject equality between WTA and WTP.
3. Discussion and Conclusion
Our results indicate that subjects respond to the prompt and revision opportunity and that
these revisions tend to be in the direction implied by the iterative valuation approach,
systematically increasing WTP and decreasing WTA.7 The process through which the prompt
would act is unclear; however, we speculate, that, coming as it does after the instructions to
iteratively arrive at one’s offer, the prompt is akin to Cheap Talk in the contingent valuation
literature—albeit a much milder form than that of the original script used by Cummings and
Taylor (1999). Using a cheap talk script is a well known method for reducing hypothetical bias
in contingent valuation surveys. A cheap talk script urges the respondent to avoid hypothetical
bias by increasing WTP to the maximum or lowering WTA to the minimum. In a similar fashion,
the prompt, in combination with the suggestion to use an iterative valuation procedure, may
suggest to subjects that they have not gone far enough in iteratively reaching maximum WTP or
minimum WTA. Importantly, in one experiment comparing hypothetical and actual payments,
cheap talk was found to over-correct for bias (Morrison and Brown, 2009). While the
effectiveness of cheap talk has been shown for some situations, there is no consensus on its
proper form or efficacy across a broad range of contingent valuation methods and researchers
have urged caution in its use (List et al., 2006; Murphy and Stevens, 2004).
In the absence of a clear benchmark from which to evaluate subjects’ offers, we cannot
say whether the pre- or post-prompt offers more accurately reflect underlying values. In light of
the evidence that the prompt does affect offers, in the implied direction, use of the prompt rests
on the assumption that some subjects will fail, prior to being prompted to revise their offers, to
reach their maximum WTP or minimum WTA, and, importantly, that subjects cannot be pushed
by suggestion beyond their true maximum WTP or minimum WTA. Another unknown is the
robustness of the prompt and revision to the experimental procedures. For example, would these
results be maintained in the absence of the iterative valuation approach described in the
instructions?
Given the lack of discussion of such prompts in the literature and the possibility that the
prompt may systematically influence offers, further research on the use of prompts is warranted.
7
As stated, our design does not completely replicate the procedures implemented by Plott and Zeiler (2005). Thus,
we cannot speculate whether a significant disparity would have been observed had Plott and Zeiler analyzed their
pre-prompt offers.
References
Binmore, K., 1999. Why experiment in economics? The Economic Journal 109 (February), F16F24.
Cummings, R.G., Taylor, L.O., 1999. Unbiased value estimates for environmental goods: A
cheap talk design for the contingent valuation method. American Economic Review 89
(3), 649-665.
Hanemann, W.M., 1991. Willingness to pay and willingness to accept: how much can they
differ? American Economic Review 81 (3), 635-647.
Horowitz, J.K., McConnell, K.E., 2002. A review of WTA/WTP studies. Journal of
Environmental Economics and Management 44 (3), 426-447.
Isoni, A., Loomes, G., Sugden, R., 2011. The willingness to pay-willingness to accept gap, the
"endowment effect," subject misconceptions, and experimental procedures for eliciting
valuations: comment. American Economic Review 101 (2), 991-1011.
Knetsch, J., 2010. Values of gains and losses: reference states and choice of measure.
Environmental and Resource Economics 46 (2), 179-188.
List, J.A., Sinha, P., Taylor, M.H., 2006. Using choice experiments to value non-market goods
and services: evidence from field experiments. Advances in Economic Analysis and
Policy 6 (2), Article 2.
Morrison, M., Brown, T.C., 2009. Testing the effectiveness of certainty scales, cheap talk, and
dissonance-minimization in reducing hypothetical bias in contingent valuation studies.
Environmental and Resource Economics 44 (3), 307-326.
Murphy, J.J., Stevens, T.H., 2004. Contingent valuation, hypothetical bias, and experiment
economics. Agricultural and Resource Economics Review 33 (2), 182-192.
Plott, C.R., Zeiler, K., 2005. The willingness to pay-willingness to accept gap, the "endowment
effect," subject misconceptions, and experimental procedures for eliciting valuations.
American Economic Review 95 (3), 530-545.
Randall, A., Stoll, J., 1980. Consumer's surplus in commodity space. American Economic
Review 70 (3), 449-455.
Willig, R., 1976. Consumer's surplus without apology. American Economic Review 66, 589-597.